Episode 12: Concurrency Pt. 1

Concurrency Basics

Alex Robson and Michael Scharf introduce the concept of concurrency, emphasizing its importance in modern computing. Concurrency allows computers to perform multiple tasks simultaneously, either through physical means like multicore CPUs or through software that simulates concurrent execution. This capability is crucial for enhancing performance and scalability in software systems. Alex explains that almost all modern operating systems support multitasking and multithreading, which are foundational to concurrency.

Concurrency can be achieved physically, which means we have more than one hardware execution unit, more than one CPU, for example, such as in modern multicore CPUs or server machines.

--- Alex Robson

This foundational understanding sets the stage for exploring more advanced concurrency topics in future episodes 1 2.

   

Processes & Threads

The distinction between processes and threads is crucial in understanding concurrency. Michael Scharf describes processes as heavyweight execution units with separate address spaces, which protect them from interfering with each other. However, this protection comes at the cost of expensive context switches and complex inter-process communication. Threads, on the other hand, are more lightweight, sharing the same address space, which allows for faster communication but increases the risk of data corruption.

Threads are more lightweight than processes. And here different threads run actually in the same address space, so they can access each other's data.

--- Michael Scharf

Understanding these differences helps developers choose the right concurrency model for their applications 3.

   

Performance & Scalability

Concurrency significantly impacts performance and scalability in computing. Alex Robson highlights that tasks requiring extensive computation, like weather simulations, benefit from concurrency by distributing workloads across multiple CPUs. This approach necessitates parallel algorithms to fully utilize available resources. Michael Scharf adds that concurrency also enhances scalability by efficiently using CPU resources, especially during I/O operations.

Regarding scalability, the idea is to split the application and put it on several CPUs, or even to use one CPU more efficiently.

--- Michael Scharf

This efficiency is vital as modern CPUs increasingly rely on parallel processing rather than higher clock speeds to improve performance 4 5.